Geospatial and XGBoost-based sub-watershed prioritization for sustainable soil and watershed management

Effective soil and water conservation planning requires identifying and ranking watersheds according to their susceptibility to soil erosion. Such prioritization helps reduce costs and improves the effectiveness of management interventions by directing resources to the most vulnerable areas. However, conventional prioritization approaches based on a single method often produce inconsistent rankings, limiting their reliability for decision-making.This study examines erosion-prone sub-watersheds within the Bikkeru watershed. The watershed was divided into eleven sub-watersheds to enable a detailed and systematic assessment. To address this limitation, a combined prioritization approach was adopted, integrating morphometric analysis, land use and land cover (LULC), and the XGBoost machine learning algorithm. The inclusion of XGBoost adds a novel dimension to the analysis by enhancing the accuracy and reliability of watershed ranking. Nineteen essential morphometric metrics were extracted from a digital elevation model using geospatial approaches. The analysis classified the Bikkeru watershed into three priority categories: high, medium, and low. Sub-watersheds SW9 and SW11 were identified as critical zones requiring immediate soil and water conservation measures, including check dams, vegetative barriers, percolation tanks, and recharge structures, to control runoff and reduce erosion. SW1, SW5, SW7, and SW10 were classified as medium priority, recommended for moderate interventions such afforestation, to reduce runoff and gully formation. SW2, SW3, SW4, SW6, and SW8 were designated as low priority zones, requiring routine monitoring and maintenance of existing conservation structures. These findings provide valuable scientific guidance for policymakers and planners in designing targeted watershed management strategies, particularly for highly vulnerable areas, and support Sustainable Development Goal 15 by enabling resource-efficient conservation planning to combat land degradation.